Radiomics strategies for risk assessment of tumour failure in head-and-neck cancer

Abstract

Quantitative extraction of high-dimensional mineable data from medical images is a process known as radiomics. Radiomics is foreseen as an essential prognostic tool for cancer risk assessment and the quantification of intratumoural heterogeneity. In this work, 1615 radiomic features (quantifying tumour image intensity, shape, texture) extracted from pre-treatment FDG-PET and CT images of 300 patients from four different cohorts were analyzed for the risk assessment of locoregional recurrences (LR) and distant metastases (DM) in head-and-neck cancer. Prediction models combining radiomic and clinical variables were constructed via random forests and imbalance-adjustment strategies using two of the four cohorts. Independent validation of the prediction and prognostic performance of the models was carried out on the other two cohorts (LR: AUC = 0.69 and CI = 0.67; DM: AUC = 0.86 and CI = 0.88). Furthermore, the results obtained via Kaplan-Meier analysis demonstrated the potential of radiomics for assessing the risk of specific tumour outcomes using multiple stratification groups. This could have important clinical impact, notably by allowing for a better personalization of chemo-radiation treatments for head-and-neck cancer patients from different risk groups.

Comparison of PET and CT radiomics for prediction of local tumor control in head and neck squamous cell carcinoma

“…It is contradictory to the study of Vallieres et al [19] in which no correlation between loco-regional control and PET/CT radiomics was found. To translate our results to loco-regional control modeling, the inclusion of additional parameters, for example lymph nodes-specific radiomics, should be considered.…”

“…Some retrospective data showed an association between disease free survival and increased maximum as well as mean standardized uptake value (SUV), whereas other groups did not observe such a correlation [18]. Recently, Vallieres et al [19] tried to predict loco-regional control using combined PET and CT radiomics. This study was not successful.…”

Personalizing Medicine Through Hybrid Imaging and Medical Big Data Analysis

“…This kind of approach successfully increased risk assessment of head-and-neck cancer built on in vivo and clinical variables with utilizing random forest ML approaches [29]. Independent cross-cohort validation revealed an AUC of 0.69 and a CI of 0.67 for predicting locoregional recurrences, while distant metastases were predicted with an AUC of 0.86 and a CI of 0.88.…”

“…Machine learning is a promising approach to deal with large-scale medical data [26]. In light of hybrid imaging, several groups have reported promising results for disease characterization by applying robust machine learning methods for combined in vivo analysis [27][28][29].…”

Prognostic value of textural indices extracted from pretherapeutic 18‐F FDG‐PET/CT in head and neck squamous cell carcinoma

“…Finally, before integrating texture analysis in overall risk stratification in HNSCC, a consensus should be reached stating which segmentation method and which resampling and calculation parameters should be applied to calculate textural indexes. Nevertheless, a recent work of the Image Biomarker Standardization Initiative and publications of Vallières et al proposed after this study design trend in this direction …”